Several projects are ongoing, designed to identify indicator foods for methodologic uses, and to describe the population's behavior with respect to potentially protective or harmful foods or nutritional supplements. Such efforts support the dietary assessments and monitoring necessary for cancer control, provide data on which to base population interventions or strategies, and shed light on factors which may influence observed subgroup differences in cancer incidence or survival. Quantitative data on nutrient sources in the U.S. diet for each of 18 nutrients have been developed. These data permit the identification of foods which make quantitatively important contributions to the population's intake of, for example, vitamin A or saturated fat. A follow-on project will provide this type of analysis for demographic or geographic subgroups. This information may aid in efforts to improve dietary assessment and dietary behavior in subgroups, as well as to understand possible reasons for observed differences in cancer incidence or survival. An analysis of the proportion of Americans consuming potentially beneficial or harmful foods, such as high-fiber foods or red meat, will be carried out for each of several categories by race, sex, age and poverty status. The results may be useful in dietary interventions and public information campaigns. Demographic and behavioral correlates of vitamin supplement usage will be described using a large representative national sample.